A Role for Interferon Regulatory Factor 4 in Receptor Editing

Abstract

Receptor editing is the primary means through which B cells revise antigen receptors and maintain central tolerance. Previous studies have demonstrated that interferon regulatory factor 4 (IRF-4) and IRF-8 promote immunoglobulin light-chain rearrangement and transcription at the pre-B stage. Here, the roles of IRF-4 and -8 in receptor editing were analyzed. Our results show that secondary rearrangement was impaired in IRF-4 but not IRF-8 mutant mice, suggesting that receptor editing is defective in the absence of IRF-4. The role of IRF-4 in receptor editing was further examined in B-cell-receptor (BCR) transgenic mice. Our results show that secondary rearrangement triggered by membrane-bound antigen was defective in the IRF-4-deficient mice. Our results further reveal that the defect in secondary rearrangement is more severe at the immunoglobulin λ locus than at the κ locus, indicating that IRF-4 is more critical for the λ rearrangement. We provide evidence demonstrating that the expression of IRF-4 in immature B cells is rapidly induced by self-antigen and that the reconstitution of IRF-4 expression in the IRF-4 mutant immature B cells promotes secondary rearrangement. Thus, our studies identify IRF-4 as a nuclear effector of a BCR signaling pathway that promotes secondary rearrangement at the immature B-cell stage.

B-cell development in the bone marrow is characterized by sequential rearrangement of immunoglobulin (Ig) heavy- and light-chain loci through a somatic DNA rearrangement event called the V(D)J rearrangement. Although the total randomness of V(D)J rearrangement is essential for the diversification of the B-cell-receptor (BCR) repertoire, it also unavoidably brings autoreactivity to the repertoire of newly generated immature B cells. Indeed, it has been estimated that 40 to 60% of newly synthesized B cells are autoreactive (29). Central tolerance is the mechanism through which developing B cells are rendered nonreactive to self. Central tolerance consists of receptor editing, anergy, and deletion (29). During receptor editing, autoreactive B cells undergo prolonged V(D)J rearrangement to replace the autoreactive heavy and/or light chain (9, 40). Anergy is a mechanism through which the autoreactive B cells are rendered inactive and, thus, unable to harm the host (10). Clonal deletion is the process through which the autoreactive B cells are depleted from the repertoire (12, 30). Recent studies have indicated that clonal deletion operates as a default pathway to get rid of autoreactive B cells that cannot be rescued by receptor editing (11, 14).

Receptor editing at the immature B-cell stage is induced by a self-reactive BCR, and it can also be induced by a BCR with an insufficient amount of tonic signaling (18). Receptor editing is a process through which self-reactive heavy or light chain is replaced with a product of secondary V(D)J rearrangement (29). Secondary rearrangement occurs mainly at the Ig κ and λ loci. The murine κ locus contains four functional Jκ elements: Jκ1, Jκ2, Jκ4, and Jκ5. During receptor editing, the primary VJκ rearrangement can be replaced by secondary rearrangement between Vκ and a downstream Jκ element. Secondary rearrangement can also occur between Vκ and a recombination sequence (RS) located ∼25 kb downstream of the Cκ or between a site located in the Jκ-Cκ intron and the RS (7). The RS rearrangement leads to functional inactivation of the whole κ locus and the initiation of Ig λ rearrangement (41).

Interferon regulatory factor 4 (IRF-4) and IRF-8 are immune system-specific transcription factors that have been shown to play critical roles in innate and adaptive immunity (39). Previous studies have demonstrated that IRF-4 and -8 function redundantly to control pre-B-cell development (21). B-cell development is blocked at the pre-B stage in mice lacking IRF-4 and -8; mutant pre-B cells are hyperproliferative and defective in light-chain rearrangement and transcription (21). Recently, we have shown that IRF-4 and -8 induce the expression of Ikaros and Aiolos to downregulate pre-BCR and inhibit pre-B-cell expansion (22). In addition, we and others have also demonstrated that IRF-4 and -8 induce chromatin modifications at the κ locus, thereby promoting κ locus activation in pre-B-cell development (20, 23). Thus, the roles of IRF-4 and -8 in pre-B-cell development are twofold: one is to limit pre-B-cell expansion and the other is to promote pre-B-cell differentiation. The molecular mechanisms through which IRF-4 and -8 control the activation of light-chain loci remain to be determined. However, previous studies have demonstrated that IRF-4 and -8 interact with Ets family transcription factors PU.1 and Spi-B to regulate the activity of the κ 3′ enhancer and λ enhancers (3, 4). In addition, IRF-4 and -8 have been found to interact with E2A to regulate the activity of the κ 3′ enhancer (27, 28).

Although the involvement of IRF-4 and -8 in light-chain rearrangement and transcription has been established in pre-B-cell development, their role in receptor editing and secondary rearrangement is still not clear. In this report, we examined the roles of IRF-4 and -8 in receptor editing. Our results show that the ratio of κ- and λ-expressing B cells is perturbed in mice deficient for IRF-4, but not for IRF-8, suggesting a unique role for IRF-4 in secondary rearrangement. Using a BCR transgenic model, we show that the secondary rearrangement triggered by membrane-bound antigen is defective in the absence of IRF-4. Moreover, we provide direct evidence demonstrating that the expression of IRF-4 in immature B cells is rapidly induced by self-antigen to promote secondary rearrangement at the light-chain loci.

MATERIALS AND METHODS

Mice.

IRF-4 mutant mice have been previously described (24). Mice expressing a transgenic BCR recognizing hen egg lysozyme (Ig^HEL) or a membrane-bound HEL antigen (mHEL) in the C57B6 background were purchased from Jackson Lab. IRF-4^−/− mice were bred with Ig^HEL mice to generate mice expressing one copy of the Ig^HEL transgene in an IRF-4-deficient background (Ig^HEL IRF-4^−/−). Ig^HEL IRF-4^−/− mice were bred with mHEL transgenic mice to generate IRF-4-deficient mice that are hemizygous for Ig^HEL and mHEL (Ig^HEL IRF-4^−/− mHEL). The mice were maintained under specific-pathogen-free conditions. Experiments were performed according to guidelines from the National Institutes of Health and with an approved IACUC protocol from the University of Nebraska Medical Center. Mice 8 to 14 weeks of age were used for this study.

FACS and cell sorting.

Cells were preincubated with either 2% rat serum or Fc-Block (2.4G2) and stained with optimal amounts of specific antibodies, either biotinylated or directly fluorophore conjugated. Antibodies against B220 (RA3-6B2), IgMa (DS-1), IgD (11-26c), and pre-BCR (SL156) were purchased from Pharmingen; anti-κ (H139-52.1) and anti-λ (JC5-1) antibodies were obtained from Southern Biotech. Fluorescence-activated cell sorter (FACS) analysis was performed with a FACSCalibur flow cytometer. To isolate immature B cells, bone marrow cells were stained with antibodies against B220 and κ and were sorted by using a BD FACSAria flow cytometer.

Culture of pre-B cells.

Pre-B cells were cultivated as described previously (23). Briefly, B220-positive (B220⁺) cells were isolated from mouse bone marrow by using a MACS separation column (Miltenyi Biotech). Purified cells were overlaid on top of an irradiated S17 stromal-cell layer. The cells were cultivated in Opti-MEM (Gibco) medium containing 5% fetal bovine serum, 50 μM β-mercaptoethanol, 2 mM l-glutamine, 100 U penicillin-streptomycin, and 5 ng/ml interleukin-7 (IL-7) (R&D). The pre-B cells were passaged every three days onto a new S17 stromal layer. Cells with fewer than five passages were used for the experiments.

Retroviral infection.

The IRF-4-expressing retroviral vector has been described previously (23). To infect primary pre-B cells, retroviral vectors containing the genes of interest were transfected into the ecotropic retroviral packaging cell line PLAT-E by using FuGene 6 (Roche). The cell-free supernatants were collected at 24 and 48 h after transfection. The virus was concentrated by centrifugation at 20,000 × g for 1 h and was typically used the same day to infect target cells via spin infection. The infection was carried out in a 24-well plate at 640 × g for 1 h in the presence of 10 μg/ml of Polybrene. The infected cells were analyzed by FACS at different time points afterwards.

Real-time PCR analysis.

The cells were lysed by using Trizol. Total RNA was extracted and reverse transcribed with a single-strand cDNA synthesis kit (Amersham). Quantitative real-time PCR analysis was carried out in an ABI 7500 real-time PCR system (Applied Biosystems) using Sybr green PCR core reagents (ABI). All samples were tested in triplicate, and average threshold cycle values were calculated and normalized to those for the housekeeping gene encoding glyceraldehyde-3-phosphate dehydrogenase (GAPDH). PCR with each primer set was independently repeated three times, and the average values and standard deviations of the results were calculated. The primers for the amplification of Vκ germ line transcripts have been described previously (8). The sequences of other primers used in this study are as follows: Rag1 for, GGAGCAAGGTAGCTTAGCCAACATGGC, and rev, CCAGGCTTCTCTGGAACTACTGGAGACTG; Rag2 for, TGTCCCTGCAGATGGTAACAGTGGG, and rev, CGAAGAGGTGGGAGGTAGCAGCAGGAATCT; Kgl for, GAGGGGGTTAAGCTTTCGC, and rev, GCCTCCACCGAACGTCCA; λ1gl for, CTTGAGAATAAAATGCATGCAAGG, and rev, TGATGGCGAAGACTTGGGCTGG; IRF-4 for, GTGGAAACACGCGGGCAAGC, and rev, GGCTCCTCTCGACCAATTCCTCA; IRF-8 for, AGAGGGAGACAAAGCTGAACCAGCC, and rev, CCACGCCCAGCTTGCATTTT; and GAPDH for, TGTGTCCGTCGTGGATCTGA, and rev, CCTGCTTCACCACCTTCTTGAT.

Intracellular Ki-67 staining and TUNEL analysis.

Splenocytes were stained with antibodies against cell surface markers, fixed, and permeabilized in 70% ethanol. The level of expression of Ki-67 was determined by using a Ki-67 staining kit (Pharmingen) according to the manufacturer's protocol. The terminal deoxynucleotidyltransferase-mediated dUTP-biotin nick end labeling (TUNEL) assay was carried out with an APO-Direct kit (Pharmingen). B cells labeled with dUTP in the absence of terminal transferase were used as the negative control. The stained cells were analyzed by FACS.

In vivo BrdU incorporation assay.

The in vivo bromodeoxyuridine (BrdU)-labeling assay was performed as described previously (1). Briefly, mice 8 to 12 weeks old were injected intraperitoneally with BrdU (Sigma-Aldrich) every 12 h for 3 days. A 200-μl amount of phosphate-buffered saline containing 0.6 mg BrdU was used for each injection. Three mice were used for each time point. Bone marrow and splenocytes were collected three days after the first injection and stained with antibodies against B220 and IgMa. The stained cells were fixed, permeabilized, and stained with anti-BrdU antibody (Pharmingen). The percentage of BrdU-positive B cells was determined by FACS analysis.

RS and λ1 rearrangement analysis.

B cells were isolated from the bone marrow and spleens of control and mutant mice via either B220-based positive selection with affinity columns (Miltenyi Biotec) or sorting with a BD FACSAria cell sorter. Genomic DNA was extracted from purified B cells after proteinase K treatment. The RS and λ1 rearrangement assay results were analyzed as previously described (2, 33). Briefly, the isolated genomic DNAs were serially diluted and used as templates for semiquantitative PCR analysis. The primer sequences for RS rearrangement were as follows: Vkcon, GGCTGCAGSTTCAGTGGCAGTGGRTCWGGRAC, and RS, CTGCCCACACGACTCCTTCAGGCAGACG; and CD19 for, AGGTAAGAAAGAGGAGGAAG, and rev, TTGTGGATTTGGAAGAGTGC. The PCR cycling conditions for amplifying the RS rearrangement product were 94°C for 45 s, 65°C for 1 min, and 72°C for 1 min for a total of 35 cycles. The primer sequences for λ1 rearrangement were as follows: λ1/2 for, AGAAGCTTGTGACTCAGGAATCTGCA, and Jλ1 rev, CAGGATCCTAGGACAGTCAGTTTGGT. The cycling conditions were 35 cycles of 94°C for 45 s, 61°C for 1 min, and 72°C for 1 min. Amplified products were resolved on a 3% agarose gel. The expression level of the CD19 gene was measured and used as a loading reference.

RESULTS

The κ and λ ratio is perturbed in the IRF-4-deficient mice.

Our previous studies have shown that IRF-4 and IRF-8 promote light-chain rearrangement and transcription in pre-B-cell development (23). Here, we sought to determine if IRF-4 and IRF-8 are also essential for receptor editing. λ-expressing B cells are often viewed as products of receptor editing and secondary rearrangement. Therefore, as an initial approach to addressing this question, we analyzed the percentages of λ-expressing IgM⁺ B cells in the bone marrow and spleens of wild-type control, IRF-4 mutant, and IRF-8 mutant mice (Fig. 1A). Our results show that the percentages of λ-expressing B cells were significantly decreased in the bone marrow and spleens of IRF-4^−/− mice. In the bone marrow, the percentage of λ-expressing B cells was 9% in control mice (IRF-4^+/+) but only 4% in IRF-4^−/− mice; in the spleen, the percentage of λ-expressing B cells was also down, from 6% in the control mice to 3% in the IRF-4^−/− mice. The percentages of λ-expressing B cells were comparable for IRF-4^+/+ and IRF-4^+/− mice, suggesting that there is no haploinsufficiency effect of IRF-4 on the κ and λ ratio. The percentage of λ-expressing B cells in IRF-8^−/− mice was comparable to the percentage in the control mice, suggesting that a lack of IRF-8 does not affect the ratio of κ- and λ-expressing B cells. The absolute numbers of B220⁺ and λ-expressing B cells in the spleens of control, IRF-4 mutant, and IRF-8 mutant mice were also enumerated (Fig. 1B and C). Although the total number of B220⁺ B cells was slightly increased in IRF-4^−/− mice, the number of λ-expressing B cells among them was only about 50% of the number in the control mice. In contrast, the number of λ-expressing B cells increased slightly in the IRF-8^−/− mice.

FIG. 1.

The κ and λ ratio was perturbed in the IRF-4-deficient mice. Cells were isolated from bone marrow and spleens of wild-type (IRF-4^+/+), IRF-4^+/−, IRF-4^−/−, and IRF-8^+/− mice. (A) The isolated cells were stained with antibodies against B220, IgM, and λ light chain and analyzed by FACS. The IgM⁺ B cells were selectively gated, and the numbers represent percentages of IgM⁺ B cells expressing λ in either the bone marrow (B.M.) or spleen. The data shown are representative of the results of at least three independent experiments. (B and C) The absolute numbers of B220⁺ B cells and λ-expressing B cells in the spleen were enumerated. The numbers are the averages and standard deviations of the results for a total of five mice in each group. (D) Secondary rearrangement is defective in IRF-4^−/− mice. Bone marrow pre-B and immature B cells (B220^low CD43^low/−) were isolated via sorting from IRF-4^+/− and IRF-4^−/− mice. Reverse transcription-PCR analysis with serially diluted templates was performed to determine RS and λ rearrangement. The CD19 genomic sequence was amplified as the loading reference.

The distorted ratio of κ and λ in IRF-4^−/− mice suggests a possible defect in secondary rearrangement. In order to confirm this finding, we measured products of secondary rearrangement, namely RS and λ rearrangement, in IRF-4^−/− and control mice. Since receptor editing in the bone marrow occurs at the pre-B and immature B stages, we isolated both pre-B and immature B cells from the bone marrow of IRF-4^+/− and IRF-4^−/− mice. As shown in Fig. 1D, the products of RS and λ1 rearrangement were significantly reduced in the IRF-4^−/− mice, indicating a defect in secondary rearrangement. Interestingly, the λ rearrangement appears to be more adversely affected than the RS rearrangement. Taken together, our results indicate that secondary rearrangement was defective in the IRF-4^−/− mice, resulting in the distorted ratio of κ- and λ-expressing B cells.

Defective generation of edited B cells in IRF-4-deficient Ig^HEL transgenic mice encountering a membrane-bound self-antigen.

It has been demonstrated that Ig^HEL B cells undergo deletion in the presence of the multivalent membrane-bound antigen HEL (12). More recently, the Ig^HEL B cells have also been shown to undergo receptor editing in the presence of membrane-bound HEL (14). In order to confirm the role of IRF-4 in receptor editing, we bred Ig^HEL transgenic mice with IRF-4^−/− mice to generate mice expressing Ig^HEL B cells in an IRF-4-deficient background (Ig^HEL IRF-4^−/−). Ig^HEL IRF-4^−/− mice were further bred with mice expressing the membrane HEL antigen to generate mice expressing Ig^HEL and membrane HEL antigen in an IRF-4-null background (Ig^HEL IRF-4^−/− mHEL). Mice that were heterozygous mutants for IRF-4 (IRF-4^+/−) were also generated and used as controls.

The B-cell development in the bone marrow and spleens of Ig^HEL IRF-4^+/−, Ig^HEL IRF-4^−/−, Ig^HEL IRF-4^+/− mHEL, and Ig^HEL IRF-4^−/− mHEL mice was examined (Fig. 2A and B). The absolute numbers of Ig^HEL B cells in the bone marrow and spleen were also counted (Table 1). The binding specificity of Ig^HEL B cells was confirmed by staining the cells with biotinylated HEL. The transgenic heavy chain is of the Ig M^a allotype (IgMa) and can be detected by staining the cells with an anti-IgMa antibody. Previous studies have shown that in the presence of the membrane-bound HEL antigen, Ig^HEL B cells in the bone marrow of Ig^HEL IRF-4^+/+ mHEL mice downregulate the surface expression of IgMa and undergo developmental arrest, resulting in an expansion of immature B cells with low levels of B220, IgMa, and HEL (B220^low IgMa^low HEL^low) (5, 12). Consistent with those findings, the expression of IgMa on immature B cells was downregulated in Ig^HEL IRF-4^+/− mHEL and Ig^HEL IRF-4^−/− mHEL mice (Fig. 2A). Moreover, compared to that in Ig^HEL IRF-4^+/− or Ig^HEL IRF-4^−/− mice, the B220^low IgMa^low HEL^low immature B-cell population also significantly increased in the Ig^HEL IRF-4^+/− mHEL and Ig^HEL IRF-4^−/− mHEL mice (Fig. 2A and Table 1). Interestingly, the number of B220^low IgMa^low HEL^low B cells in the Ig^HEL IRF-4^−/− mHEL mice was significantly higher than that in Ig^HEL IRF-4^+/− mHEL mice, suggesting that the lack of IRF-4 expression may exacerbate the developmental arrest caused by the membrane-bound self-antigen (Fig. 2A).

FIG. 2.

Defective generation of edited B cells in Ig^HEL IRF-4^−/− mice encountering a membrane-bound self-antigen. (A and B) Cells were isolated from bone marrow and spleens of Ig^HEL IRF-4^+/−, Ig^HEL IRF-4^−/−, Ig^HEL IRF-4^+/− mHEL, and Ig^HEL IRF-4^−/− mHEL mice at 8 to 10 weeks of age. Cells were stained with indicated antibodies and analyzed by FACS. The numbers indicate the percentage of cells that fall into each quadrant. (C) The splenic B cells were isolated from Ig^HEL IRF-4^+/− mHEL and Ig^HEL IRF-4^−/− mHEL mice 14 to 15 weeks of age. The cells were stained with antibodies against IgMa, κ, and λ and analyzed by FACS. The results shown are representative of at least three independent experiments.

TABLE 1.

B-cell populations in the bone marrow and spleen^a

Phenotype of mice	No. of B cells (105) in:
	Bone marrow		Spleen
	HEL+	HEL−	HEL+	HEL−
IgHEL IRF-4+/+	13.2 ± 4.2	8.3 ± 2.6	125.7 ± 13.5	8.8 ± 1.0
IgHEL IRF-4+/−	10.9 ± 5.4	12.1 ± 6.0	107.3 ± 10.1	9.2 ± 0.9
IgHEL IRF-4−/−	15.5 ± 6.5	10.1 ± 4.2	71.9 ± 19.8	12.4 ± 3.4
IgHEL IRF-4+/+ mHEL	38.1 ± 7.3	10.1 ± 1.9	1.9 ± 0.5	23.3 ± 6.1
IgHEL IRF-4+/− mHEL	37.3 ± 10.4	14.6 ± 4.1	2.3 ± 0.7	19.1 ± 5.8
IgHEL IRF-4−/− mHEL	59.2 ± 14.6	14.8 ± 3.6	3.2 ± 1.5	4.3 ± 2.0

^aThe HEL⁺ and HEL⁻ B cells were counted separately. The numbers are the means ± standard deviations of the results for a total of 6 mice in each group.

It has been demonstrated that in the presence of membrane-bound antigen, the majority of splenic Ig^HEL B cells were deleted, with the exception of those that had replaced the transgenic light chain with an endogenously rearranged light chain, thereby losing the reactivity to HEL (12). Indeed, compared to the number in Ig^HEL IRF-4^+/− mice, the total number of B cells was dramatically reduced in the spleens of Ig^HEL IRF-4^+/− mHEL mice (Fig. 2B and Table 1). Consistent with the results of previous reports, the majority of splenic B cells in Ig^HEL IRF-4^+/− mHEL mice did not recognize HEL (Fig. 2B). However, the majority of the splenic B cells still expressed IgMa, indicating that the failure to bind to HEL is due to a replacement of transgenic light chain with an endogenous light chain as a result of receptor editing. Strikingly, about 50% of the splenic B cells in Ig^HEL IRF-4^−/− mHEL mice still recognized HEL, albeit with a low affinity (Fig. 2B). The edited splenic B cells (IgMa⁺ HEL⁻) in the Ig^HEL IRF-4^+/− mHEL mice were mature B cells expressing a high level of B220 (B220^hi), whereas the splenic B cells in Ig^HEL IRF-4^−/− mHEL mice were B220^low, suggesting that, phenotypically, these cells were still immature B cells. In addition, the total number of splenic B cells in Ig^HEL IRF-4^−/− mHEL mice was approximately one-third of that found in Ig^HEL IRF-4^+/− mHEL mice (Table 1).

To determine if the edited B cells could be generated in older IRF-4-deficient mice, splenic B cells in Ig^HEL IRF-4^−/− mice of 14 to 15 weeks of age were analyzed (Fig. 2C). Interestingly, in addition to the B220^low HEL⁺ population, we were able to detect a population of B220^hi HEL-edited B cells in the Ig^HEL IRF-4^−/− mHEL mice. However, the edited population was present at approximately one-fifth of the number found in the Ig^HEL IRF-4^+/− mHEL control mice (Fig. 2C). These results suggest that the edited B cells may be generated at a slower rate in the IRF-4-deficient background. We further analyzed the expression of λ light chain in the edited B cells. Our results show that about 8% of the splenic B cells in Ig^HEL IRF-4^+/− mHEL mice expressed λ light chain, whereas the splenic B cells in the Ig^HEL IRF-4^−/− mHEL mice expressed almost exclusively κ light chain (Fig. 2C). Taken together, these results suggest that edited B cells are generated at a slow rate in IRF4-deficient mice.

Secondary rearrangement activity is defective in the Ig^HEL IRF-4^−/− mHEL mice.

The slow generation of the edited B cells in Ig^HEL IRF-4^−/− mHEL mice could be a result of impaired receptor editing in the absence of IRF-4. To examine this possibility, we decided to measure secondary rearrangement activity in B cells isolated from the bone marrow and spleens of Ig^HEL IRF-4^+/−, Ig^HEL IRF-4^+/− mHEL, and Ig^HEL IRF-4^−/− mHEL mice. Products of RS and λ rearrangement in the isolated bone marrow and splenic B cells were examined by PCR. We isolated DNA from mice at 8 weeks as well as 14 weeks of age in order to determine the possible effect of age on receptor editing. RS and λ1 rearrangement could not be detected in DNA isolated from the bone marrow and spleens of Ig^HEL IRF-4^+/− mice at 8 weeks of age, indicating that secondary rearrangement activity was very low in those mice (Fig. 3). However, in the presence of membrane-bound antigen, RS and λ1 rearrangement could be readily detected in B cells isolated from the bone marrow and spleens of Ig^HEL IRF-4^+/− mHEL mice, indicating that the secondary rearrangement activity was significantly increased as a result of receptor editing. Compared to the results for Ig^HEL IRF4^+/− mHEL mice, the product of RS rearrangement was diminished, whereas λ1 rearrangement could not be detected in the Ig^HEL IRF-4^−/− mHEL mice at 8 weeks of age, suggesting that the secondary rearrangement activity was impaired in the IRF-4-deficient mice (Fig. 3). The secondary rearrangement activity remained impaired in Ig^HEL IRF-4^−/− mHEL mice at 14 weeks of age. However, the differences in RS rearrangement between Ig^HEL IRF-4^+/− mHEL and Ig^HEL IRF-4^−/− mHEL mice were less dramatic than those seen in the younger mice, suggesting that the defect in secondary rearrangement, particularly RS deletion, could be partially compensated in the older Ig^HEL IRF-4^−/− mHEL mice. In summary, our results show that secondary rearrangement was impaired in the Ig^HEL IRF-4^−/− mHEL mice.

FIG. 3.

Secondary rearrangement activity was defective in the Ig^HEL IRF-4^−/− mHEL mice. Cells were isolated from the bone marrow and spleens of Ig^HEL IRF-4^+/−, Ig^HEL IRF-4^+/− mHEL, and Ig^HEL IRF-4^−/− mHEL mice and were subjected to positive selection to purify B220⁺ B cells. Genomic DNA was extracted from the B220⁺ cells and analyzed by PCR to detect RS and λ1 rearrangement. Mice of 8 or 14 weeks of age were used and analyzed separately. Templates were serially diluted to allow semiquantitative analysis. A fragment of the CD19 gene was amplified and used as an internal loading reference. The results shown are representative of three independent experiments.

Impaired secondary rearrangement in IRF-4-deficient mice is not caused by defects in cell proliferation or apoptosis.

V(D)J rearrangement activity is cell cycle dependent, and it takes place at the G₀/G₁ stage of the cell cycle (6). Our recent studies have demonstrated that IRF-4 induces the expression of Ikaros and Aiolos to inhibit pre-B-cell expansion (22). Therefore, it is possible that in the absence of IRF-4, immature B cells may exhibit a higher proliferation index which would indirectly inhibit secondary rearrangement activity. To determine the cell cycle status of the cells, we measured the levels of expression of Ki-67, an intracellular protein found only in cycling cells. The majority of the B220^low IgMa⁺ immature B cells in the bone marrow of Ig^HEL IRF-4^+/− and Ig^HEL IRF-4^−/− mice stained negative for Ki-67, indicating that the majority of these cells were quiescent (Fig. 4A). In contrast, approximately 50% of the B220^low IgMa⁻ cells (pro-/pre-B) still expressed Ki-67, indicating that they were cycling. In the presence of membrane-bound antigen, 7% of IgMa⁺ B cells expressed Ki-67 in both the Ig^HEL IRF-4^+/− mHEL and Ig^HEL IRF-4^−/− mHEL mice, indicating that the defect in secondary rearrangement in Ig^HEL IRF-4^−/− mHEL mice was not due to an increased proliferation index.

FIG. 4.

Impairment of secondary rearrangement activity in IRF-4-deficient mice was not caused by defects in cell proliferation or apoptosis. (A) Cells were isolated from the bone marrow of Ig^HEL IRF-4^+/−, Ig^HEL IRF-4^−/−, Ig^HEL IRF-4^+/− mHEL, and Ig^HEL IRF-4^−/− mHEL mice, stained with antibodies against B220 and IgMa, fixed, and permeabilized in 70% ethanol. The permeabilized cells were stained with anti-Ki-67 antibody and analyzed by FACS. IgM⁺ and IgM⁻ cells were gated separately, and the numbers represent percentages of TUNEL-positive B cells. (B) Cells were isolated from the bone marrow (B.M.) and spleens of Ig^HEL IRF-4^+/−, Ig^HEL IRF-4^−/−, Ig^HEL IRF-4^+/− mHEL, and Ig^HEL IRF-4^−/− mHEL mice, stained with biotinylated HEL and anti-B220 antibody, fixed, and permeabilized. The percentages of TUNEL-positive cells were determined by FACS. The numbers are the means and standard deviations of the results for a total of three mice in each group. The data are representative of the results of three independent experiments. *, P value of <0.05 in comparison to the results for Ig^HEL IRF-4^+/− mHEL mice. (C) In vivo BrdU-labeling assay. Mice of 8 to 12 weeks of age were injected intraperitoneally with BrdU every 12 h for 3 days. Bone marrow and splenocytes were collected and stained with antibodies against B220, IgMa, and BrdU antibodies. Three mice were used for each time point. The percentages of BrdU-positive bone marrow (gated on IgMa⁺ B cells) and splenic (gated on B220⁺ cells) B cells were determined by FACS analysis. The numbers are the averages and standard deviations of the results for each group. *, P value of <0.05 in comparison to the results for Ig^HEL IRF-4^+/− mHEL mice.

An increase in apoptosis in IRF-4-deficient B cells can also indirectly affect secondary rearrangement and reduce the number of the edited B cells in the Ig^HEL IRF-4^−/− mHEL mice (19). To determine if there is a defect in apoptosis, a TUNEL assay was conducted to measure the percentages of B cells undergoing apoptosis (Fig. 4B). The HEL⁺ and HEL-negative (HEL⁻) B cells were analyzed separately in the spleen. The percentage of TUNEL-positive B cells was higher in the Ig^HEL IRF-4^+/− mHEL and Ig^HEL IRF-4^−/− mHEL mice than in the Ig^HEL IRF-4^+/− and Ig^HEL IRF-4^−/− mice, suggesting that B cells may undergo apoptosis at an increased rate in the presence of self-antigen (Fig. 4B). The results of the TUNEL analysis also reveal that in Ig^HEL IRF-4^+/− mHEL and Ig^HEL IRF-4^−/− mHEL mice, similar percentages of immature B cells undergoing receptor editing were apoptotic, indicating that a lack of IRF-4 did not lead to significantly enhanced apoptosis in the antigen-activated immature B cells (Fig. 4B). Similarly, the percentage of apoptotic B cells was similar for the edited splenic B cells (B220⁺ HEL⁻) in Ig^HEL IRF-4^+/− mHEL and Ig^HEL IRF-4^−/− mHEL mice. However, compared to the results for Ig^HEL IRF4^+/− mHEL mice, the percentage of apoptotic cells was moderately increased among the population of HEL⁺ splenic B cells in the Ig^HEL IRF-4^−/− mHEL mice (P < 0.05).

An increase in B-cell apoptosis can lead to an increase in the B-cell turnover rate. To further examine the status of B-cell apoptosis, the turnover rate of B cells in the IRF-4-proficient and -deficient mice was examined after three days of labeling with BrdU (Fig. 4C). Our results show that in the bone marrow, the percentage of BrdU-labeled IgM⁺ B cells was comparable for Ig^HEL IRF-4^+/− mHEL and Ig^HEL IRF-4^−/− mHEL mice. However, in the spleen, 38% of B cells in Ig^HEL IRF-4^−/− mHEL mice were labeled with BrdU compared to 25% in Ig^HEL IRF-4^+/− mHEL mice, indicating an enhanced B-cell turnover rate in the spleens of Ig^HEL IRF-4^−/− mHEL mice. Nevertheless, since B cells undergoing receptor editing in the bone marrow showed comparable turnover rates and exhibited similar levels of apoptosis in the IRF-4-proficient and -deficient backgrounds, apoptosis is not likely to have been a major contributor to the defective secondary rearrangement observed in the Ig^HEL IRF-4^−/− mHEL mice.

Expression of IRF-4 is rapidly induced when immature B cells encounter self-antigen to promote secondary rearrangement at the κ and the λ loci.

Our results show that secondary rearrangement activity was defective in the Ig^HEL IRF-4^−/− mHEL mice. We sought to determine the molecular mechanism by which IRF-4 regulates secondary rearrangement. To this end, we isolated and compared the gene expression profiles of bone marrow B cells undergoing receptor editing (B220^low IgMa⁺) in the Ig^HEL IRF-4^+/− mHEL and Ig^HEL IRF-4^−/− mHEL mice. It has been shown that immature B cells undergoing receptor editing maintain the expression of Rag1 and Rag2. Indeed, compared to their levels in the immature B cells isolated from Ig^HEL IRF-4^+/− mice, the expression levels of Rag1 and Rag2 were significantly elevated in B cells isolated from both Ig^HEL IRF-4^+/− mHEL and Ig^HEL IRF-4^−/− mHEL mice (Fig. 5A). These results also indicate that the defect in secondary rearrangement in the Ig^HEL IRF-4^−/− mHEL mice was not due to deregulated Rag1 and Rag2 expression.

FIG. 5.

IRF-4 expression in immature B cells is rapidly induced by self-antigen to promote secondary rearrangement at the κ and λ loci. (A) Bone marrow immature B cells (B220^low κ⁺) were isolated via sorting from Ig^HEL IRF-4^+/−, Ig^HEL IRF-4^+/− mHEL, and Ig^HEL IRF-4^−/− mHEL mice. Real-time PCR was performed to measure the expression of the indicated genes. kGL, κ germ line transcript; l1GL, λ1 germ line transcript. *, P < 0.05; **, P < 0.01. P values are in comparison to the results for the Ig^HEL IRF-4^+/− mHEL mice. (B) Immature B cells were isolated via sorting from the bone marrow of Ig^HEL IRF-4^+/+ mice. The sorted cells were treated with HEL at 400 ng/ml and lysed at different time points for RNA extraction. The levels of expression of IRF-4 and IRF-8 were measured by real-time PCR. The numbers are the averages and standard deviations of the results for each group. (C) B220⁺ B cells were isolated from the bone marrow of Ig^HEL IRF-4^−/− mice and expanded in culture in the presence of IL-7 (5 ng/ml). The cells were infected with retrovirus expressing either control or IRF-4. IL-7 was removed from the medium 36 h later, and the cells were incubated with HEL for three more days. The infected cells (green fluorescent protein positive) were isolated via sorting and were subjected to DNA extraction. PCR analysis was performed to measure RS and λ1 rearrangement. The expression of CD19 was used as a loading reference. (D) The expression of IRF-4 was measured by Western blot analysis in the infected cells and the cultured Ig^HEL IRF-4^+/+ immature B cells. Actin was used as a loading control. Data are representative of the results of three independent experiments.

The expression of light-chain germ line transcripts has been associated with the activation of the light-chain loci (36). To determine if there is a defect in the activation of light-chain loci in the absence of IRF-4, the expression levels of κ and λ1 germ line transcripts were analyzed (Fig. 5A). Our results show that κ germ line transcript in B cells isolated from Ig^HEL IRF-4^+/− mHEL mice was expressed at a level that was approximately twofold higher than the level in Ig^HEL IRF-4^−/− mHEL mice, whereas λ1 germ line transcript expression was approximately fivefold higher than that in Ig^HEL IRF-4^−/− mHEL mice (Fig. 5A). These results suggest that there was a defect in the activation of light-chain loci in the Ig^HEL IRF-4^−/− mHEL mice. We further measured the germ line transcripts of four major Vκ families: hf24, Gn33, 12-38, and 21-3. Consistent with the expression patterns of κ and λ1 germ line transcripts, three of the four Vκ germ line transcripts were also expressed at significantly lower levels in the IRF-4-deficient mice (Fig. 5A). Taken together, our results indicate that there was a defect in the activation of the light-chain loci in the IRF-4-deficient mice.

The levels of expression of IRF-4 and IRF-8 in isolated immature B cells were also examined (Fig. 5A). Our results show that IRF-4 was expressed at a level that was about fivefold higher in the Ig^HEL IRF-4^+/− mHEL mice than in the Ig^HEL IRF-4^+/− mice, suggesting that the expression of IRF-4 may be induced by antigen in the immature B cells. In contrast, the expression of IRF-8 was slightly decreased in immature B cells isolated from the Ig^HEL IRF-4^+/− mHEL and Ig^HEL IRF-4^−/− mHEL mice (Fig. 5A).

In order to determine if the expression of IRF-4 is induced by self-antigen at the immature B-cell stage, immature B cells were isolated via sorting from the bone marrow of Ig^HEL IRF-4^+/+ mice and activated in vitro by treatment with HEL antigen. Our results show that the expression of IRF-4 was rapidly induced in the presence of HEL antigen; the induction could be detected as early as 2 h after the addition of HEL (Fig. 5B). In contrast, the expression of IRF-8 was not induced by HEL. These results indicate that the expression of IRF-4 is rapidly induced by antigen in immature B cells.

We further sought to determine if reconstituting the expression of IRF-4 in Ig^HEL IRF-4^−/− immature B cells would promote secondary rearrangement in the presence of self-antigen. To this end, B220⁺ B cells were isolated from the bone marrow of Ig^HEL IRF-4^−/− mice and cultivated in the presence of IL-7. Retroviral infection was performed to reconstitute IRF-4 expression. The infected cells were treated with HEL for four days and isolated via sorting to extract genomic DNA. Compared to that in the control infected cells, secondary rearrangement activity was dramatically increased in the presence of IRF-4, as evidenced by elevated RS and λ1 rearrangement (Fig. 5C). The results of Western blot analysis indicate that in the infected Ig^HEL IRF-4^−/− immature B cells, IRF-4 was expressed at a level that was comparable to that in the Ig^HEL IRF-4^+/+ immature B cells (Fig. 5D). Taken together, our results suggest that the expression of IRF-4 is induced by antigen in immature B cells to promote secondary rearrangement.

DISCUSSION

A series of recent studies have demonstrated critical functions for IRF-4 at several stages of B-cell development. It has been shown that IRF-4 is an essential transcriptional regulator for pre-B-cell development, germinal center reaction, class-switch recombination, and plasma-cell differentiation (17, 21, 22, 35, 38). The results of our studies presented here provide evidence that IRF-4 is also important for receptor editing at the immature B-cell stage, suggesting that IRF-4 is critical for the induction of B-cell tolerance. Our results show that secondary rearrangement was defective in the IRF-4-deficient wild-type mice, as well as in the Ig^HEL transgenic mice, in the presence of membrane-bound antigen. Our studies further reveal that the impairment of secondary rearrangement in the IRF-4-deficient mice was the result of a defect in the activation of light-chain loci. The κ and λ germ line transcripts in the Ig^HEL IRF-4^−/− mHEL mice were expressed at lower levels than in the Ig^HEL IRF-4^+/− mHEL mice, indicating a defect in the activation of κ and λ loci in the absence of IRF-4. Ig κ and λ loci are sequentially activated during B-cell development: κ rearrangement almost always occurs prior to λ rearrangement (25, 32, 34). λ rearrangement is induced by a self-reactive BCR and, thus, is often viewed as a product of receptor editing and secondary rearrangement. Interestingly, in the absence of IRF-4, the expression of λ germ line transcript appeared to be more adversely affected than that of the κ, suggesting that IRF-4 is more critical for the activation of the λ locus. Consistent with this view, our results show that the defect in secondary arrangement was more severe at the λ locus than at the κ locus in the Ig^HEL IRF-4^−/− mHEL mice; edited splenic B cells consisted of both κ- and λ-expressing cells in the Ig^HEL IRF-4^+/− mHEL mice but contained almost exclusively κ-expressing cells in the IRF-4-deficient background. Taken together, our results suggest that the differential dependence on IRF-4 for their activation could be the molecular basis for the sequential activation of κ and λ loci in B-cell development.

Previously, we have shown that IRF-4 and IRF-8 function redundantly to promote κ locus activation in pre-B-cell development (23). However, the defects in RS and λ rearrangements in Ig^HEL IRF-4^−/− mHEL mice suggest that IRF-4, but not IRF-8, played a dominant role in secondary rearrangement. Consistent with this view, we show that IRF-4 expression, but not IRF-8 expression, is induced in immature B cells undergoing receptor editing in vivo. Moreover, the expression of IRF-4, but not IRF-8, is rapidly induced when immature B cells encounter self-antigen. However, our finding that secondary rearrangement, though impaired, could still be detected in the Ig^HEL IRF-4^−/− mHEL mice also indicates that B cells are still able to revise the antigen receptor in the absence of IRF-4. Furthermore, our results show that the defects in secondary rearrangement, particularly RS rearrangement, could be partially compensated in the older Ig^HEL IRF-4^−/− mHEL mice. One possible explanation for these observations is that secondary rearrangement is less efficient in the absence of IRF-4, and, thus, may require a longer time for successful rearrangement of the endogenous light-chain gene. Consistent with this idea, our results show that, compared to that in Ig^HEL IRF-4^+/− mHEL mice, the population of B cells undergoing editing was significantly enlarged in the bone marrow of Ig^HEL IRF-4^−/− mHEL mice. Collectively, our results support a scenario in which IRF-8 may play a role in maintaining the basal activity of the light-chain loci, particularly at the κ locus, whereas the elevated expression of IRF-4 in the presence of self-antigen leads to further activation of the κ and the λ loci, thereby promoting efficient secondary rearrangement. Our finding that reconstituting IRF-4 expression in the IRF-4-deficient immature B cells promotes secondary rearrangement is also consistent with this view.

Receptor editing can be influenced by changes in cell survival and apoptosis status. It has been shown that enhanced survival prolongs the life span of immature B cells, thereby promoting receptor editing, whereas reduced cell survival may indirectly inhibit receptor editing (19). Thus, reduced survival of the immature B cells in the absence of IRF-4 would be sufficient to suppress secondary rearrangement. However, our results do not reveal a significant difference in B-cell apoptosis in the bone marrow of IRF-4-proficient and -deficient mice, as the percentages of TUNEL-positive immature B cells, as well as the turnover rate of the immature B cells, were similar in IRF-4-proficient and -deficient mice. In addition, our finding that Ig^HEL IRF-4^−/− mHEL mice had a much larger immature B-cell pool than Ig^HEL IRF-4^+/− mHEL mice also does not support the notion that immature B cells undergo elevated apoptosis in the IRF-4-deficient mice. However, our findings do reveal that the splenic B cells in the Ig^HEL IRF-4^−/− mHEL mice underwent elevated apoptosis compared to that of their counterparts in the Ig^HEL IRF-4^+/− mHEL mice. Since significant numbers of splenic B cells still recognized HEL in Ig^HEL IRF-4^−/− mHEL mice, the increased numbers of the apoptotic splenic B cells could be a result of enhanced deletion due to inefficient receptor editing in the IRF-4-deficient mice. Another possible explanation for these data could be that IRF-4, though not essential for the survival of immature B cells, may be an important survival factor for the peripheral B cells. Therefore, it is possible that inefficient receptor editing and the enhanced apoptosis may both have contributed to the slow generation of the edited splenic B cells in the Ig^HEL IRF-4^−/− mHEL mice.

BCR signaling triggered by self-antigen at the immature B-cell stage not only induces the expression of Rag1 and Rag2 but could also lead to further activation of the light-chain loci to promote efficient secondary rearrangement (31). It has been demonstrated that the BCR-induced expression of Rag1 and Rag2 at the immature B-cell stage is dependent on transcription factor NF-κB (42). The NF-κB family of transcription factors has also been implicated in the demethylation of the light-chain κ locus and, thus, may play a role in κ locus activation (15). Previous studies have demonstrated that transcription factor E2A is critical for light-chain rearrangement at the pre-B stage and for receptor editing at the immature B-cell stage (16, 33). It has also been demonstrated that IRF-4 can recruit E2A to the Ig κ 3′ enhancer in pre-B cells (20, 27). Therefore, it is possible that elevated IRF-4 in the immature B cells can interact with and recruit E2A to the Ig κ locus to promote receptor editing. In contrast, the activation of the Ig λ locus by IRF-4 is most likely mediated by interaction with the Ets family transcription factors PU.1 and Spi-B (3, 4); it has been shown that λ, but not κ, transcription is severely impaired in the PU.1 and Spi-B double-deficient pre-B cells, indicating that the interaction between IRF-4 and PU.1/Spi-B is critical for λ, but not κ, locus activation (37). The expression of IRF-4 is induced at the pre-B stage to promote light-chain rearrangement and transcription (23, 26). Although the results of our studies favor the scenario where elevated IRF-4 in the autoreactive immature B cells promotes/maintains the activation of light-chain loci for efficient receptor editing, we cannot rule out the possibility that the defective activation of light-chain loci in the IRF-4-deficient autoreactive immature B cells is partially caused by the lack of IRF4 at the pre-B stage of their development.

Receptor editing is initiated when immature B cells encounter self-antigen. The importance of BCR signaling in receptor editing has been illustrated by findings that mutations of the components of the BCR signaling pathway lead to defects in secondary rearrangement (2, 13). The results of our studies presented here identify IRF-4 as a potential nuclear effector of a BCR-initiated signaling pathway that promotes secondary rearrangement at the immature B-cell stage.